This is a pure object database. The objects actually stored need to implement a base class.
*It stores objects in clusters. It sounds like originally a cluster was a group of related objects. But now the cluster has evolved into a "xml document" where each element is an object.
*It can be used to store xml documents.

It is very similar to what the LabBook looked liked for teemss1.

I haven't figured out how it will handle synchronization: Do I need to instanciate every object inorder to do a copy part of the tree?

http://www.ozone-db.org

There are some interesting related projects:

• TM4J - a topic map, I have yet to figure out what this is but my initial intuition is that it is like the activity map that we are using in pedagogica.